Abstract Pedestrian detection is crucial for various applications, including intelligent transportation and video surveillance systems. Although recent research has advanced pedestrian models like the YOLO series, they still face limitations in handling diverse scales, leading to performance challenges. To address these issues, we propose HF-YOLO, an model. HF-YOLO tackles complexities of complex scenes by addressing scale variations occlusions among pedestrians. In feature fusion stage, our...

Real-time and accurate traffic forecasting aids in planning design helps to alleviate congestion. Addressing the negative impacts of partial data loss forecasting, challenge simultaneously capturing short-term fluctuations long-term trends, this paper presents a model, D-MGDCN-CLSTM, based on Multi-Graph Gated Dilated Convolution Conv-LSTM. The model uses DTWN algorithm fill missing data. To better capture dual characteristics trends traffic, employs DWT for multi-scale decomposition obtain...

Sustainable ecosystem services consumption is of vital importance to the survival and development human society. How balance conflicts between protection by local residents has been a serious challenge, especially in ecologically vulnerable areas. To explore reasonable approaches grassland ecosystems for sustainable land system management, this study takes Hulun Buir Inner Mongolia Autonomous Region as case region develops an EcoC-G (ecological grassland) model based on herders’ livelihood...

Farmland is the most basic material condition for guaranteeing rural livelihoods and national food security, exploring management strategies that take both stable sustainable farmland use into account has vital significance in theory practice. a complex self-adaptive system couples human natural systems, social factors are related to its changing process need be considered when modeling processes. This paper uses Qianjingou Town Inner Mongolian farming–pastoral zone as study area. From...

Purpose This paper aims to propose a new ultrasonic diagnosis method for stainless steel weld defects based on multi-domain feature fusion solve two problems in the of austenitic defects. These are insufficient extraction and subjective dependence model parameters. Design/methodology/approach To express richness one-dimensional (1D) signal information, 1D testing was derived two-dimensional (2D) time-frequency domain. Multi-scale depthwise separable convolution also designed optimize...

Open-vocabulary learning has recently gained prominence as a means to enable image segmentation for arbitrary categories based on textual descriptions. This advancement extended the applicability of systems broader range generally purpose scenarios. However, current methods often revolve around specialized architectures and parameters tailored specific tasks, resulting in fragmented landscape models. In response these challenges, we introduce OVAMTSeg, versatile framework designed...

An image segmentation approach based on independent component analysis and watershed algorithm is proposed to complement medical efficiently. Considering that the noise, uniform gray intensity many low contrast adjacent areas in images may cause over-segmented when using algorithm, so applied reduce noise' affection image-segmentation keep clear texture simultaneously, then used for image-segmentation. Compared method wavelet filter classic experiments display this efficient segmentation.

Urban livability is an important factor in decisions regarding modern urban construction. Based on multi-agent modeling theory, a model of the changes safety proposed to analyze present situation and trends complex systems. Considering as example, five environmental indicators, namely public security, fire protection, traffic, emergency response, residence, are selected. The interactive dynamic feedback mechanisms among resident decision-making behaviors context these indicators relevant...

Urban livability is an important factor affecting the sustainable development of modern cities. Safe part urban livability. In view this, this paper takes security as example, and based on actual data Futian District, Shenzhen City, China, establishes a multi-agent simulation model safe The dynamic interaction feedback mechanism between decision-making behaviors residents under influence environment policies has been explored, residents’ change realized. Finally, main factors influencing are...

Improving the urban livability status has become core goal of development, and reasonable assessment impact is crucial. By combining an objective environment with residents’ subjective cognition, artificial society (urban change society; ULC-AS) constructed. The ULC-AS includes four types agents, namely, government, family, resident safety facility management agency recognizes dynamic interaction among various agents between environment. Taking Futian District Shenzhen as example, this paper...

The security of car driving is interest due to the growing number motor vehicles and frequent occurrence road traffic accidents, combination advanced assistance system (ADAS) vehicle-road cooperation can prevent more than 90% accidents. Lane detection, as a vital part ADAS, has poor real-time performance accuracy in multiple scenarios, such damage, light changes, jams. Moreover, sparse pixels lane lines on pose tremendous challenge task line detection. In this study, we propose model that...

Incomplete diagnostic information, inadequate multisource sensor weak diagnosis models, and subjective experience result in difficulty predicting rotating machinery faults. To overcome these limitations, we proposed a multiple domain heterogeneous information entropy fusion model based on an optimisation of bearing fault diagnosis. The spatiotemporal approach uses multiscene strategy sensors (HSMSF) to extract feature strategies analyses the characteristics features by multichannel processes...

Abstract We propose a simple image high-perception super-resolution reconstruction method based on multi-layer feature fusion model with adaptive compression and parameter tuning. The aim is to further balance the high low-frequency information of an image, enrich detailed texture improve perceptual quality, optimization generalization in process reconstruction. First, effective (MFSR) basic constructed by design edge enhancement, refine layering, enhanced generative adversarial network...

In this paper, an original workflow is presented for MR image plane alignment based on tracking in real-time images. A test device consisting of two resonant micro-coils and a passive marker proposed detection using image-based algorithms. Micro-coils allow automated initialization the object dedicated low flip angle projection images; then tracked clinical images, with alternation between oblique orthogonal planes along axis; case lost reinitialized. The was designed to minimize acquisition...